FIELD: image data processing.
SUBSTANCE: invention relates to the field of computer graphics and can be used for synthesising a talking head, animating a face. According to the proposed approach, the appearance of a person is modelled by decomposing into two layers. The first layer is a pose-dependent rough image synthesised by a relatively small neural network. The second layer is determined by a pose-independent texture image containing high-frequency details. The texture image is created in the off-line mode and distorted, and then added to the rough image in order to provide high effective resolution of the synthesised views of the head.
EFFECT: increase in the resolution and increase in the operation speed. The proposed system creates neural avatars based on a single photograph.
10 cl, 2 dwg
Title | Year | Author | Number |
---|---|---|---|
TEXTURED NEURAL AVATARS | 2019 |
|
RU2713695C1 |
METHOD FOR 3D RECONSTRUCTION OF A HUMAN HEAD TO OBTAIN A RENDER IMAGE OF A PERSON | 2022 |
|
RU2786362C1 |
METHOD AND SYSTEM FOR REMOTE CLOTHING SELECTION | 2020 |
|
RU2805003C2 |
MODELLING PEOPLE'S CLOTHING BASED ON MULTIPLE POINTS | 2021 |
|
RU2776825C1 |
NEURAL-NETWORK RENDERING OF THREE-DIMENSIONAL HUMAN AVATARS | 2021 |
|
RU2775825C1 |
NEURAL NETWORK TRANSFER OF THE FACIAL EXPRESSION AND POSITION OF THE HEAD USING HIDDEN POSITION DESCRIPTORS | 2020 |
|
RU2755396C1 |
IMAGE GENERATORS WITH CONDITIONALLY INDEPENDENT PIXEL SYNTHESIS | 2021 |
|
RU2770132C1 |
REPEATED SYNTHESIS OF IMAGE USING DIRECT DEFORMATION OF IMAGE, PASS DISCRIMINATOR AND COORDINATE-BASED REMODELLING | 2019 |
|
RU2726160C1 |
METHOD OF CREATING FULL-LENGTH ANIMATED AVATAR OF PERSON FROM ONE IMAGE OF PERSON, COMPUTING DEVICE AND MACHINE-READABLE MEDIUM FOR IMPLEMENTATION THEREOF | 2023 |
|
RU2813485C1 |
SYSTEM AND METHOD FOR OBTAINING PROCESSED OUTPUT IMAGE HAVING USER-SELECTABLE QUALITY FACTOR | 2023 |
|
RU2823750C1 |
Authors
Dates
2022-01-13—Published
2020-07-27—Filed